[Tweeters] Merlin Complaints Backfire ! Badly.
Chuq Von Rospach via Tweeters
tweeters at u.washington.edu
Mon Feb 2 09:46:18 PST 2026
On Feb 2, 2026 at 01:03:31, Robert O'Brien via Tweeters <
tweeters at u.washington.edu> wrote:
> For quite a while, there have been complaints that some bird calls/songs
> are reported as birds that 'cannot possibly be present'.by Merlin. Merlin
> is often held up to scorn.
> But what if the bird IS ACTUALLY present, as a vagrant. For instance
> rare eastern warblers breeding/singing in coastal SW Oregon
>
My take is that if the bird is that rare and your only ID data is a merlin
ID, that’s not a good enough ID to add it to eBird, so I think Merlin is
doing the right thing here. Without stronger evidence (visual confirmation,
enough experience with the call to call it without assistance), that’s not
a solid ID in my mind so it lives on as a “something”.
But for experienced birders, that means vagrants are out of bounds for
> Merlin. Either location or seasonal wise.
>
I think this is the right call. Newer birders tend to trust Merlin
explicitly (ask any eBird reviewer). Experienced birders will many times
(but not always) know enough to be skeptical, and experienced enough to not
make an eBird catch definitive without confirming evidence. I’m all for
eBird protecting the naive here.
That said, there are some things Merlin ought to do that it’s not that
annoy me.
First, I experimented for a while with a tool called Haiku, which is one of
a number automatic song pickup systems you can get now. It’s pretty good in
a number of ways, but among other things, having a microphone listening
outside my door 24x7 just wasn’t my idea of fun. One thing I want Merlin to
add is Haiku’s confidence rating — it not only gives you the ID, but a
oonfidence level (high/medium/low) for that ID. That way, you can have some
sense of how good the ID is, and Merlin makes it seem all of its IDs are
absolutely perfect (which it’s not)
Second, why doesn’t Merlin make it easy to extract a given call for further
analysis or uploading with the eBird report? This should be fairly easy to
implement, either on the fly or as a later “re-process” phase offline, and
would make it a lot easier to do follow up listening or sharing with others
for these less likely IDs. Right now, the recording stays one big mass of
data that you have to manually pull apart and honestly, that’s too much
hassle for almost all IDs. They could make improving this and adding good
call data to eBird a lot simpler.
Overall though, while I have my issues with Merlin, I think it’s gotten way
better in the last few years, and I’m all for it leaning towards doing the
right thing for newbies here.
Chuq
---------------------------------------
Chuq Von Rospach (http://www.chuq.me)
Silverdale, Washington
Birder, Nature and Wildlife Photographer
Email me at: chuqvr at gmail.com
Mastodon: @chuqvr at fosstodon.org
Stay Updated with what I'm doing: https://www.chuq.me/6fps/
My latest e-book: https://www.chuq.me/ebooks
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://mailman11.u.washington.edu/pipermail/tweeters/attachments/20260202/ad986bcb/attachment.html>
More information about the Tweeters
mailing list